A game turns an image into a place where actions have consequences. In this part of my 2024 research, I consider how AI could make those places more responsive and easier to create. These are prospects for visual culture and game design, not a claim that every game is becoming a world without limits.
Procedural generation and AI are not the same thing, although they could be combined. Rules can generate a landscape; AI tools might help develop its textures, inhabitants or variations. The design task would be to give these elements a coherent visual language. A larger world would not necessarily be a more memorable one: a distinctive place may matter more than endless terrain.
Non-player characters could respond to a wider range of actions and conversations. Their apparent intentions might make an encounter less predictable, while changes in posture, movement or expression communicate what is happening. Yet a convincing character would still be a designed system. Consistent behaviour and understandable limits could matter more than an unlimited ability to produce dialogue.
Stories might branch around a player's decisions, with environments showing the consequences: a changed room, an altered route or a place revisited in a different state. AI could help explore these variations. Meaning would still depend on rhythm, memory and relationships between events. Generating another plot development is different from making the player care about it.
Avatars and personal spaces could become more adaptable. Players might choose an appearance, rearrange a setting or request a different presentation of information. Expressed preferences would be a clearer starting point than assuming a system can read emotions. In shared worlds, personalisation would also need common rules so that people can understand one another's actions and play fairly.
Production tools could help small teams prototype, prepare assets and test particular situations. This might make more visual approaches feasible, without making a finished game effortless. Editing, performance, rights to source material and testing would remain part of the work. Easier generation could support independent voices—or spread the same familiar aesthetic through many games.
I see the possibility of games becoming a more accessible medium for constructing experiences from our visual inheritance. The important question would be what an author chooses to make playable. A responsive world would be valuable not because it produces something new at every moment, but because the player's choices leave a meaningful trace.

AI & Visual Culture: 39. Theatre ↗

